Eindhoven University of Technology
MASTER
Fire behavior of sandwich panel core materials in the pre-flashover phase
Giunta d'Albani, A.W.
Award date:
2014
Link to publication
Disclaimer
This document contains a student thesis (bachelor's or master's), as authored by a student at Eindhoven University of Technology. Student theses are made available in the TU/e repository upon obtaining the required degree. The grade received is not published on the document as presented in the repository. The required complexity or quality of research of student theses may vary by program, and the required minimum study period may vary in duration.
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
COLOPHON
Fire Behavior Of Sandwich Panel Core Materials In The Pre-flashover Phase
A study into sandwich panels and their fire safety properties, based on a literature study resulting in experiments. To gain further insight in possible hazards of steel insulated sandwich panels in the pre-flashover phase.
For the graduation of the master track Building technology, of the department Architecture, Building and Planning at the Eindhoven University of Technology.
By: Ing. A.W. Giunta d’Albani 0773711
Tutored by:
Prof. Dr. Ir. H.J.H. Brouwers TU/e Ir. R.A.P. van Herpen FIFireE Ir. A.C.J. de Korte, TU/e
Dr. Ir. R. Weewer (Brandweer Nederland)
In cooperation with:
Nederlands Instituut voor Fysieke Veiligheid (IFV) Brandweer Nederland
Technische Universiteit Eindhoven(TU/e) Edition: Definitief
Version 1
4-11-2014
At the start of this graduation project, my knowledge about fire safety and insulation materials was limited.
Since my study has its main focus on building technology, material science is a new direction, nevertheless a challenging one. The knowledge about these new subjects has been gained by reading literature, attending a Fire safety engineering course, a National Fire safety engineering congress organized by the IFV, and the National Fire safety congress organized by SBRCUnet. Besides the congresses we have visited two production plants (Isobouw at Someren and Kingspan at Tiel), where meetings and presentations gave insight in the world of synthetic insulation materials. Conversations with different consultants from the company’s, Peutz an DGMR have helped to gain knowledge on the field fire safety tests.
This indicative research has been done under the supervision of Prof. Dr. Ir. H.J.H Brouwers (TU/e), Ir. R.A.P van Herpen FIFireE, Dr. Ir. R. Weewer (Brandweer Nederland), and Ir. A.C.J. de Korte of the Technical University Eindhoven. I would like to thank them for their time and knowledge. And I would like to thank the entire scientific Board of Brandweer Nederland for their enthusiasm and input.
Besides the support of the supervisors I would like to give our special thanks to several experts in the field of fire safety, materials and fire tests to start with Roy Weghorst, and Benedikt van Roosmalen for their continuous input, useful remarks and pleasant and helpful meetings. I also like to thank Peter van de Leur (DGMR) for his comments on the literature study, and a refreshing view of the test setup and the instruments to measure these experiments. As well as Jacques Mertens (Peutz) who helped to design the small experimental setup.
Preface
Abstract
This research is commissioned by the Dutch fire department and the Eindhoven University of Technology (TU/e). The Dutch fire department is increasing their scientific knowledge of events that are happening during a fire. They suspect sandwich panels with a combustible core to be a potential hazard, after several serious events in buildings constructed with these panels. This research contains a summary of the literature study and an indicative research. The literature study done together with L.L de Kluiver gives a short overview of the available literature about fire hazards regarding sandwich panels in an objective way. The research made use of different journals in the field of fire safety engineering and building materials, case study reports made by different companies, as well as information from manufacturers and Euro codes. Topics involved in this research are: sandwich panel, thin metal sheets, combustible insulation, polyurethane, polyisocyanurate, pyrolysis, and steel facings. The indicative research mainly contains experiments and simulations in order to create insight in the total amount of pyrolysis products in a smoke layer.
Both parts of the research are to check the validity of the concern from the Dutch fire department.
The results from the literature study show that the concern of the Dutch fire department might be correct although most of the literatures focuses on fully developed fires. The indicative research is designed to create more insight in the actual behavior in the pre-flashover phase. In order to determine if the sandwich panels, which are mainly applied on Dutch buildings pose, threats for fire fighters
The literature study has focused on the hazards of the sandwich panel as a building product. The fire hazards of combustible cored sandwich panels are not clear yet. Fire fighters see things happening which are not supposed to happen according to official fire tests.
Literature describes delamination and falling down of the metal facings, which are a potential hazards for fire fighters. Official fire tests do not describe this event. Besides that, cores can be exposed earlier in real fires than fire test show, due to a greater influence of the buckling effect. The mixture of smoke gases and pyrolysis gases can become a potential mix for smoke-gas explosions.
The results of the indicative research show: that delamination does occur in the pre-flashover phase, all tested samples show signs of delamination or loss of structural strength at the upper range of the
pre-flashover phase. Panels are normally mounted on purlins, so the chance of panels falling down is minimal. Some deformation of panels has been seen, but this test has been to small to make any statement about the buckling effect. Surprisingly the actual mass- loss by pyrolysis of synthetic cores and mineral wool based cores does not differ much up to 300 degrees Celsius. The mass-loss of PUR panels is exponential and PUR starts losing a significant amount of mass around 300 degrees Celcius, whereas the PIR and stone wool panels have a lower pyrolysis temperature and show a more linear trend. The radiation flux of the smokelayer is an important factor by which the maximum temperature can be calculated. The fire fighting tenability limits has been suggested as a maximum radiation flux of 4,5 Kw/m² at 1.5 m above the floor and a minimum height to the bottom of the smoke layer of 2 m” [19]. A temperature of 350 °C would give a heat flux of 8 Kw/m², this radiation level marks the upper limit for a fire fighters to apply an offensive fire repression within the fire compartment.
Since the radiation level can be reduced by a factor 0.5 when working close to the sides of the compartments.
When results of the experimental test are used in the Spreadsheet calculations. The simulations show that even in the most extreme situations the limit of the 39% mass of flammable gases of the total smoke layer will not be reached. The poultry farms (long and low buildings) show with 22.6% the highest amount of flammable gasses of the total smoke layer. Most buildings do not generate more than 5% pyrolysis gases of the total smoke layer at temperatures up to 270 degrees Celsius ( 4,5 Kw/m²), and 11% at temperatures up to 350 degrees Celcius ( 8 Kw/m²) During these simulations completely closed buildings where calculated. The outlet of pyrolysis gases has not been taken in account. All results generated in these calculations are project specific, and dot not contain compartment larger than 1600m².
To conclude, sandwich panels with synthetic and mineral wool cores both emit pyrolysis gases when exposed to heat. The potential hazard of a flammable mixture occurring in the smoke layer lies above the temperature range to which fire fighter can safely be exposed. Sandwich panels in any form are not dangerous as a building material, and do not pose any threat during normal use of the building. However, during a fire the presence of combustible additives in sandwich panels may in some cases, and mainly after pre-flashover increase the intensity of the fire.
beschrijft de delaminatie van de stalen bekleding en zelfs het vallen van deze bekleding, en het vervroegd blootstellen van kernen door vervorming en of het doorbuigen van panelen. En daarnaast het mogelijk ontstaan van rooklaag explosies door het uitgassen van de panelen
De resultaten van het indicatieve onderzoek laten vet volgende zien: delamimantie dan wel verlies van draagkracht treed op in de pre-flashoverfase, dit wordt waargenomen aan de bovengrens waaraan brandweerlieden blootgesteld kunnen worden.
Daarnaast worden sandwichpanelen gewoonlijk ondersteund door gordingen. Het vallen van de stalen bekleding van staal sandwichpanelen als dak bekleding is dan ook hoogst onwaarschijnlijk.
Lichtelijke vervorming van de proefstukken is waargenomen tijdens de uitgevoerde testen, maar de test is te kleinschalig om hier betrouwbare conclusies uit te trekken. In geen enkele situatie wordt de grens van 39 % brandbare gassen in de rooklaag overschreden.
De meeste gebouwen generenen niet meer dan 5%
pyrolise gassen, bij een rooklaag temperatuur tot 270 graden Celsius ( 4.5 Kw/m²), en 11% bij 350 graden Celcius (8 Kw/m²). De maximum temperatuur waaraan een brandweerman/vrouw blootgesteld kan worden, wordt mede bepaald door de stralingsflux.
Als limiet wordt een straling van 4.5 Kw/m² op een hoogte van 1. 5 meter boven vloer niveau met een rookvrije hoogte van 2 meter gesuggereerd [19].Er zijn geen openingen in de simulaties mee genomen waardoor er geen afvoer van rook plaatsvind, men kan hier dus spreken van een worst-case scenario.
Een ander verrassend resultaat van dit onderzoek is dat het massaverlies van synthetische kernen en mineralewolkernen niet ver uiteenlopen tot een temperatuur van 300 graden.
Concluderend: sandwichpanelen met synthetische- en mineralewolkernen stoten beide gassen uit wanneer deze blootgesteld worden aan hitte. Het mogelijke gevaar van een brandbaar mengsel in de rooklaag, veroorzaakt door alleen de sandwich panelen ligt boven de temperatuurgrens waaraan brandweerlieden blootgesteld kunnen worden. Sandwichpanelen in welke vorm dan ook vormen geen gevaar voor personen bij normaal gebruik van het gebouw.
De aanwezigheid van brandbare toevoegingen in de sandwichpanelen kan in sommige gevallen, en met name na de pre-flashoverfase bijdragen aan de intensiteit van de brand.
Dit onderzoek is uitgevoerd in opdracht van Brandweer Nederland en de Technische Universiteit Eindhoven (TU/e). Brandweer Nederland is bezig met het vergroten van hun wetenschappelijke kennis over fenomenen die optreden ten tijde van brand. Brandweer Nederland vermoedt dat sandwichpanelen met een synthetische kern mogelijk gevaarlijke situatie s opleveren ten tijde van een brand. Dat vermoeden is aangewakkerd naar aanleiding van enkele ernstige branden in gebouwen waarin vermoed word dat synthetische isolatie en rol speelt bij het ontstaan van een brandbaar mengsel in de rooklaag, en de intensiteit van de brand.
Dit onderzoek bevat een samenvatting van het literatuuronderzoek, aangevuld met een indicatief onderzoek. Het literatuuronderzoek gedaan in samenwerking met L.L. de Kluiver geeft een kort overzicht van de beschikbare literatuur over de gevaren van sandwichpanelen. Door gebruik te maken van verschillende wetenschappelijk tijdschriften op het gebied van fire safety engineering en materiaal wetenschappen, en informatie verkregen van producenten. Aandachtsgebieden van het onderzoek zijn: sandwichpanelen, dunne staalplaten, brandbare isolatie, polyurethaan, polyisocyanuraat, pyrolyse, stalen bekleding.
Het indicatieve onderzoek bevat de resultaten van de uitgevoerde experimenten, simulaties, en berekeningen om de mogelijke risico’s van het ontstaan van een brandbaar rookgas mengsel in de rooklaag te benaderen. Beide onderdelen van het onderzoek zijn gedaan om het vermoeden van Brandweer Nederland te bevestigen dan wel te ontkrachten. Het resultaat van het literatuuronderzoek wijst erop dat het vermoeden van Brandweer Nederland correct is. De gebruikte literatuur focust zich op volledig ontwikkelde branden.
Het indicatieve onderzoek is opgezet om meer inzicht te creëren in het gedrag van de sandwichpanelen in de pre- flashoverfase. Om te kunnen concluderen of de meest gangbare sandwichpanelen toegepast in Nederland mogelijk risico’s vormen voor brandweerlieden ten tijde van een offensieve binnen-inzet.
Het literatuuronderzoek heeft zich gericht op de gevaren van sandwichpanelen als bouwproduct. De gevaren van de panelen ten tijde van een brand zijn nog niet duidelijk vastgelegd. Brandweerlieden zien dingen gebeuren welke volgens officiële brandtesten niet zouden moeten gebeuren. De gevonden literatuur
Samenvatting
Inhoud
COLOPHON 3 Preface 5 Abstract 6 Samenvatting 7 Glossary 10
1. Introduction 13
1.1 General introduction 13
1.2 Reading guide 13
1.3 Sandwich panels in Literature 14
1.3.1 The sandwich panel 14
1.3.2 Polyurethane foam (PUR) 14 1.3.3 Polyisocyanurate foam (PIR) 14
1.3.4 Mineral wool 14
1.3.5 Ignition temperatures and char 15 1.3.6 Fire test and objections 15
1.4 Aim and scope 17
1.5 Research questions 17
1.6 Limitations 18
2. Experimental Test Setup 20
2.1Test-procedure 20
2.2 Test setup 21
2.3 Used instruments 22
2.4 Future possibilities test setup 23
2.5 Validation test setup 24
2.5.1Test-runs 24
2.5.2 Sample placing procedure 25 2.5.3. Behaviour in outside conditions 25
3. Results Experiments 26
3.1 Test evaluation 26
3.2 Test results PUR 26
3.3 Test results PIR 28
3.4 Test results stone wool roof (SWR) 29 3.5 Test results stone wool wall (SWW) 30
3.6 Overall analysis 31
3.7 Comparing to the literature 33
4. Simulations 36 4.1 Indication by Comsol Multiphysics® 36 4.1.2 Results Comsol Multiphysics® 36 4.1.3 Validation Comsol Multiphysics® 38
4.2 Indication Voltra 39
4.2.1 Results Voltra 40
4.2.2 Validation Voltra 40
4.3 Conclusion 41
4.4 Ozone/ spreadsheet calculations 41
4.4.1 Ozone 41
4.4.2 Excel 41
4.4.3 Types of buildings 42
4.4.4 Results mass-loss spreadsheet calculations 44 4.4.5 Analysis spreadsheet calculations 46
4.5 Radiation flux 47
5. Conclusion, discussion and advise 48
5.1 Conclusion 48
5.2 Possible solutions to reduce the risks of
flammable mixtures 49
5.3 Limitations/discussion 49
References 50
Figure index 51
Appendix A: Temperature validation 56 Appendix B: Voltra simulation report PIR 61 Appendix C: Voltra simulation report PIR 68 Appendix D: Voltra simulation report SWR 75 Appendix E: Excel simulations summary 82
Appendix F: Sample pictures 90
Appendix G: Sample weight 94
Glossary
Buckling:
The event that happens due to thermal stress and loss of load bearing capacity.
Delamination:
Separation of different layers applied in sandwich panels. Caused by loss of strength of the adhesives that connect the facings to the cores.
EN:European standards, valid for all European country’s.
EN-ISO:
ISO Standard accepted by the standardization institute for the European union. Valid in Europe.
Explosion:
An exothermic chemical process which, when it happens at a constant volume, generates a sudden and significant increase in pressure.
Facings:
Thin metal sheets or other materials applied on the insulation cores of the sandwich panels. As protection of the core, with hygienic and aesthetic meaning.
Flammability range:
This is the range within which gas/air mixtures can ignite.
Flash ignition:
The temperature, on which a material can be ignited with a flame or spark.
Flashover:
A flashover is a transition period from when the fire is burning locally until the whole room is involved in the fire. A flashover occurs when the room conditions exceeds a particular critical level. Factors contributing to the increase in the fire’s heat release rate, include flame spread over combustible surfaces and reradiation from the hot smoke gas layer. A flashover marks the transition from the early fire development stage to a fully developed compartment fire.
Flame spread:
The speed in which the fire travels on and trough the material.
Fuel controlled fire:
After ignition and at the start of a fire’s development, the fire is described as fuel controlled as there is sufficient air for combustion and the fire’s development is controlled entirely by the fuel’s properties and arrangement. A fire can also be fuel controlled at a later stage in its development.
Heat release rate:
When a material combusts it releases heat. This released heat is measured in watts (J/s).
Influenced surface:
The area of sandwich panels that has been exposed to the smoke layer.
Intruding depth:
The depth on which the thermal degradation is visually reconcilable after exposing the samples to the heat source.
ISO:International organization for standardization.
This organization manages, and develops standards on a lot of different fields and products, on a world wide base. Only valid when accepted by a country or European union.
LCA-score:
Life Cycle Assessment of Building Assemblies and Materials.
LFL:Lower flammability limit, the lowest ratio of gases and air that can be ignited.
LPCB:
Loss prevention certificate board. An organization that tests materials on their fire safety properties, and attach these to a certificate.
Mass-loss:
The difference in total mass before and after testing the sample.
Mass-loss rate:
The mass-loss rate is the speed at which pyrolysis occurs from a material, sometimes also known as the rate of pyrolysis. This is measured in g/m2s.
NEN:NEderlandse Norm, Manages The Dutch standardization process, comparable with ISO but on a national level. Valid in the Netherlands.
NEN-ISO:
ISO standard accepted in The Netherlands by the dutch standardization institute (NEN). Also valid in The Netherlands.
NEN-EN-ISO:
A code for a standard accepted by The Dutch Standardization Institute and the European institute for standardization. Valid in the European Union.
PIR:Polyisocyanurate (insulation material).
Pre-flashover:
The fase in which the fire develops, from the very beginning of the fire till it reaches flashover.
PUR:Polyurethane (insulation material).
Pyrolysis:
Pyrolysis is a chemical decomposition process or other chemical conversion from complex to simpler constituents, caused by the effect of heat.
Pyrolysis material smoke layer ratio:
This ratio shows the amount of pyrolysis gas as a percentage of the total smoke layer ratio.
Self-ignition:
The moment on which a material has reached such a high temperature that it will ignite by itself.
Sandwich panel:
A triple layered self-supporting building insulation element, mainly used for fast construction. And available in different types and materials. Often existing out of a non-combustible core and facings.
Sample temperature:
The temperature as measured in the sample while exposing it to the heat source.
Smoke gas explosion:
When unburned smoke gases leak into an area adjacent to the fire room, they can mix very well with air to produce a combustible mixture. If there is an ignition source available or one becomes available some other way, the smoke gases can ignite with an extremely devastating effect. As a rule, this phenomenon occurs seldom.
Surface temperature:
The temperature of the inner steel facing while being exposed to the heat source.
UFL:Upper flammability limit, the limit where the ratio of flammable gases is so high that when exceeded this mixture can not be ignited.
Ventilation controlled fire:
As the fire grows it may become ventilation controlled when there is no longer sufficient oxygen to combust the pyrolysis gases formed. The fire’s heat release rate is then controlled completely by the amount of air which is available, in which case the fire is described as being ventilation controlled.
XPS:Extruded polystyrene (insulation material).
1 Introduction
1.1 General introduction
In contemporary buildings in which we live and work, higher and higher insulation standards are to be achieved in order to save energy. This aim for higher insulation levels has introduced widespread of synthetic and wool based insulation materials.
The five most commonly used insulation materials are the polyurethane foams, polyisocyanurate foams, expanded/extruded polystyrene, stone wool and glass wool based products. Generally called synthetic insulation materials or combustible insulation materials and mineral wool based insulation materials or non-combustible insulation materials. Since the emergence of the synthetic insulation materials, their behavior in case of a fire has been questioned.
Although synthetic insulation manufacturers market their products as fire safe, without extra health risk when in a fire situation. It is commonly known and proved that mineral wool are supposed to have Excellent fire safety properties. Manufacturers of mineral wool based insulation guarantee a fire safe product with a high insulation level. This division of all insulation materials into synthetic and mineral wool based insulation materials, may cause a troubled overview of the actual fire behavior of the individual products. Since there is a lot of diversity in the properties of the insulation materials, an indicative research into those materials to determine the actual behavior in the event of a fire is needed.
This research is initiated by the Dutch fire department to determine the possible risk that fire fighters are exposed to, during their offensive fire repression in buildings on which steel insulated sandwich panels are applied. The Dutch fire brigade has seen some serious events in buildings on which synthetic insulation materials are applied. They suspect that these materials increase the intensity of the fire, and might possibly cause smoke layer explosions in the pre-flashover phase. Producers of synthetic insulation materials claim otherwise. This shows the urgency and need for an objective and indicative research into the synthetic and mineral wool based insulation materials.
1.2 Reading guide
This study is an experimental follow-up on the study, Fire behavior of synthetic insulation materials in sandwich panels, by L.L. de Kluiver and A.W. Giunta d’ Albani. In this research, the first part contains the aim of the mass-loss experiments, and their experimental testsetup. Followed by the explanation of the spreadsheet calculations, which will be used to answer the question wether steel insulated sandwich panels pose any threats to fire fighters active in offensive fire repression inside a building, regarding flammable mixtures in smoke layers
The second chapter of this report contains the technical explanation of the setup followed by the validation of the test setup. On the field of temperature behavior within the furnace.
The third chapter contains the analysis of the results of the experimental setup (mass-loss) with detailed information about relative mass-loss and degradation of each sample type.
Chapter four includes the result of all the simulations.
Temperature prediction through the core materials as well as smoke layer calculations in large building compartments.
Final chapter five will end with a conclusion about the fire safety issues of steel insulated sandwich panels in the pre-flashover phase.
1.3 Sandwich panels in Literature
This chapter contains some basic information about sandwich panels and the different types of cores applied in sandwich panels. For every type,the relevant fire properties are stated. Finally, the standard fire tests and associated fire class system will be explained. Not every available type of sandwich panel is discussed, only the most common types are elaborated.
1.3.1 The sandwich panel
Sandwich panels are a popular building product due to: their weight, the thermal insulation capacity, span, dimensions, and a fast assembly process during construction. Combined with the large diversity of the aesthetic variety, high LCA-score, and the relatively low impact on the environment during the production process. These properties make combustible-cored sandwich panels not only financially attractive, but also attractive on the field of sustainability. Therefore, sandwich panels are used in many different building envelopes. At first they where applied in low life- risk building, such as storages or factories. “Due to the improvement of their aesthetic and hygienic properties, and the increase of variety, nowadays sandwich panels are applied in schools, hospitals, prisons, retails outlets, and other public buildings.”[1]
Since the field in which sandwich panels are applied has grown, the chance of them being exposed to a fire has increased automatically.
A steel insulated sandwich panel exists of a core that has a high insulation capacity, wrapped in thin steel or aluminium facings, but other materials like wood, plastics or paper are also possible. This research will only focus on the panels with metal facings. The facings are attached to the core by adhesives. There are two main categories of sandwich panels cores, the non-combustible core panels, such as glass-wool and stone-wool and the combustible core (EPS, XPS, PUR, PIR) panels. In order of frequency of use, sandwich panels for external roof and wall applications are:
PUR, PIR, LPCB approved PIR, mineral wool (rock fibre), EPS, and mineral wool (glass fibre).[2]. One can conclude that the most common cores in sandwich panels applied in The Netherlands are made out of the combustible cores: PIR and PUR.
PIR/PUR steel sandwich panels, also known as ruggedized foam panels, are produced according the Rigid Faced Double Belt Lamination (RF-DBL) process. Here the foam will be connected to the
steel facings during the expansion process of the foam in order to create one solid element. Since this is a continuous process the element can be cut to the desired length. This production process is only applied on panels with thin metal facings.
When the right materials are applied, sandwich panels can reach Excellent fire resistance. Each core has its own properties as described below.
1.3.2 Polyurethane foam (PUR)
Polyurethane is an organic insulation material made from a reactive mixture of two principal liquid components and a number of additives to produce a foam with a closed cell structure. The foam produced will not normally be ignited by a small heat source but a larger flame will cause ignition and flame spread, resulting in abundant toxic smoke production. Since PUR is a thermoset it will not melt but it will pyrolyse.
The pyrolysis temperature is 200oC. The flash ignition temperature is 320oC-420oC. The self ignition is 420oC-550oC with a calorific value of approximately 26 MJ/Kg [3][4]. The achievable fire resistance, according to the LPS 1208 for PUR is approximately 30 minutes[2]. This is shown in Figure 2.1.
1.3.3 Polyisocyanurate foam (PIR)
PIR is also a thermoset and is produced in the same way as PUR, but the ratio between the components and types of additives is usually different, to produce a polymer with a crosslinked closed cell structure, and higher fire resistance. The process control factors are of higher importance compared to PUR. Decomposion while being exposed to heat is slowed down due to the formation of a char layer. The higher fire resistance is mainly obtained by the production of char, which creates a thin layer of protection. This process can be compared to the char forming, on for example wood.
PIR has a calorific value of approximately 24 MJ/Kg [3]. The achievable fire resistance according to the LPS 1208 for PIR is approximately 30 minutes except LPCB approved PIR which is fire resistance till 35 minutes, as shown in Figure 1.1[2].
1.3.4 Mineral wool
Mineral wool products, such as glass wool and stone wool, are inorganic products. The amount of binders used in low density stone wool is negligible. Although the binders are used to improve the density and strength. Meaning that a higher density stone wool panel will contain more binders, these binders and the glue used to make stone wool sandwich panels
will pyrolyse, when exposed to heat. The calorific value of the core will vary for each type of sandwich panel. Due to the binders used in stone wool, the fire load of stone wool is comparable to EPS respectively 714 MJ/M³ for stone wool and 792 MJ/M³.[5]
According to NEN 6090, the calorific value of stone wool insulation needs to be taken in account while defining the maximum compartment size.
1.3.5 Ignition temperatures and char
As can be seen in Figure 1.2 on the next page there are no specific ignition temperatures of the materials.
Intervals have been taken since exact temperatures of ignition depend on the composition[6]. These differences in temperatures can be explained by the diversity in density and chemical additions. An increase of density means an increase of fire resistance.
Rigid insulation foams are more easily available to a fire than wood, although they do not have much different ignition temperatures, due to the low density of the foam. Wood will last longer during a fire due to its higher density, and the creation of a char layer.
The use of a char layer to delay the breakdown of the panel core and increase the fire resistance is applied in PIR foams. Therefore the flash ignition temperature has increased approximately 40 degrees compared to PUR foams. The flash ignition temperature of EPS is a bit lower than PUR, A loss of structural strength for sandwich panels can be expected around 200oC- 300oC for PIR and PUR panels due to pyrolysis.
1.3.6 Fire test and objections
Since February 2002, the European Union has set European standards for reaction to fire as the primary classification system for testing building products.
This system has an advantage to the old systems, while it tests all building products with the same tests in all countries according to one standard. For testing fire resistance there are six classes: A1, A2, B, C, D, E and F in which the products will be classified. The A1 class is the highest fire resistance class of the Euro code. An A1 classified product is a non-combustible product and does not contribute to a fire in any way.
A product that has a high fire load or low ignition temperatures will be classified as a class E material.
If the material is not tested or fails the small flame test, it automatically is a class F product. When a product is tested it is, CE approved (Conformité Européenne). Since the first of January 2003, all used building products applied in buildings, have to be CE certified. This classification is done by four different tests: Single Burning Item test (SBI-test), calorific bomb, small flame test, and the ISO furnace test.
The ISO- furnace test (EN ISO 1182): is the test method
where products will be tested for the classifications A1 and A2. For 60 minutes small samples will be exposed at a temperature of 750 degrees Celsius. This test is also known as the incombustibility test.
The calorific bomb test is used to determine the calorific value. This is done by placing a sample under high pressure (30Bar) with pure oxygen in a hermetically sealed cylinder or bomb. This cylinder or bomb is placed in water as specified in EN ISO 1716. The sample will be ignited and the calorific value is determined by the rise of temperature.
Figure 1.1: Fire resistance achievable for sandwich panels Taken from Fire Resistance of Sandwich Panel Systems[2].
Figure 1.2: Graph of the different ignition temperatures based on the gas temperature of fire room
Note graph 2.2: All temperatures shown in the graph, show temperatures of small samples of core materials. Temperatures of sandwich panels as a package may diverge from the temperatures shown in this graph due to the protection of the steel facings and the low thermal conductivity.
The small flame test as described in the EN ISO 11925-2 is used for classifications in the classes B, C, D and E. It is used to determine the capability of ignition and flame spread of a product. This is done by igniting the sample, using a candle light which is placed underneath the sample. The temperature above the flame is 180 oC from this point. The sample will be marked each two centimeter till a height of 15 centimeter. The flame is not allowed to cross the 15 centimeter mark during the test.
The Single Burning Item test (EN 13823) is designed for the product classifications A2, B, C and D.
Sandwich panels will always be subjected to the SBI test even when the core material is classified as A1 product due to the adhesives used in the panels (ISO 1182). This test simulates a starting fire. The sample will be exposed for 20 minutes to a flame of 30 KW. During this test oxygen use, smoke and CO2 production is measured. The used oxygen is in proportion to the power of the fire. This test does not
Figure 1.3: Fire safety classes as stated in the EUROCODE Taken from the EUROCODE Council DIrective 89/106/EEC [87]
include the influence of smoke gases to expansion of the fire. The classification has been regulated in the Euro-code. Figure 1.3 shows the terms of each class [7][8].
There have been several reports respectively, [1][9]
[10][11][12]that conclude that the SBI test is not representative for fires in real buildings. According to prof. Cooke (2004)[9] and Prof van Hees (2005) [13], full scale test are required to determine the actual fire safety of sandwich panels. Larger samples will show more distortion of panel faces and the buckling effect of the panels will be larger. “When the spans are bigger than 4.5 meters this buckling can cause panels falling from the ceiling” [14]. The supporting construction/purlins will prevent the facings from falling down as prof. Cooke describes. The buckling and panel face distortion are of great influence to the fire resistance properties of the product. Since they will expose the core of the sandwich panels to the fire. “The heat exposure and the sample size of the
SBI test are too small to cause any severe buckling of the panels. Therefore most panels with a combustible core will obtain the highest possible classification in the SBI test which essentially means that they should not cause flash over in a small room. However, many of those panels will go to flash over when tested in The ISO 9705.”[15]. The CEN has developed large scale tests (EN 14837) especially for steel sandwich panels with combustible cores to overcome the objections to the SBI test.
1.4 Aim and scope
The aim of this research is first of all to gain insight in the possible dangers for firefighters who enter a building on which sandwich panels are applied. And secondly to find answers to the unrequited questions around insulations materials exposed to heat in the pre-flashover fire conditions. It should be clear at the end of this research whether synthetic insulation materials are as dangerous, and wether stone wool based insulation materials are as incombustible as believed. The main focus lies on the pre-flashover phase, which is a relatively low temperature range (100-500 °C) in the field of fire and related material engineering. This means that the behavior of materials in the pre-flashover phase is different compared to higher temperatures. Based on the literature as described before, it is suspected that all materials will stay in the pyrolysis phase. Which makes it interesting since this phase means that products will not ignite but will emit possibly flammable gases. Pyrolysis gases combined with these low temperatures make a possible accumulation of flammable pyrolysis gases in the smoke layer plausible. The right amount of flammable pyrolysis gases can create a dangerous situation such as smoke layer explosions.
Insulation materials come in many different varieties, properties and products, this variety of products, comes with a variety of ingredients and recipes to produce the insulation materials. This research will focus on the insulation materials applied in steel sandwich panels, which are used in the Netherlands, more specifically panels with a Polyurethane, Polyisocyanurate and stone wool core. The stone wool panels are taken in account as reference material. The tested panels will not involve all core materials applied in buildings, EPS steel sandwich panels have been produced in the past, due to technical and financial reasons these panels are not produced for new building projects. 0,25-0,5% of the applied flat roof constructions is still in use. This is a relative small amount. A research into EPS panels is, therefore not as relevant as PIR/PUR panels.
Stone wool panels also have a low market share in
the Dutch market due to moist problems, weight and cost. It remains interesting to examine the difference between these panels and combustible cored panels.
The total market for steel insulated sandwich panels is relatively small, since most industrial building are built using the steel deck principle. But the pyrolysis range, combined with the simplicity of the product makes a research into steel insulated sandwich panels interesting to start with.
1.5 Research questions
The main research question for this research is:
Does the mass-loss of a sandwich panel by pyrolysis produce enough gases to reach the flammability limits when exposed to temperatures between 100-400 degrees?
Which has been divided into two sub research questions. The first one to be answered by the developed experimental setup and the second one to be answered by spreadsheet calculations:
What visible changes occur when exposing sandwich panels to heat and what is their mass-loss?
This question will result into a comparison between the stone wool based cores and the synthetic cores.
To see how these materials behave when exposed to temperatures as occur in the pre-flashover phase.
Results will be generated by the experimental mass- loss test, in which the sample will be exposed to a constant temperature for a constant time. There has been chosen to develop a test setup since the current available test are not designed to test the behavior of sandwich panels as a product in the pre-flashover phase. Fire resistance tests, which qualify the fire safety classes of a product, expose the sample to temperatures far above flashover phase. The mass- loss of the core materials can also be determined by a TGA analysis, which will give some insight in the mass-loss related to the temperature. The influences of a sandwich panel as a construction are not taken in account. In the developed test which will be explained in chapter 2, the sandwich panel as a product will be exposed to a heat source/hot air, which simulates a smoke layer. In this test the mass- loss will be determined, regarding the influence of the steel facings:
What are the expected concentrations of pyrolysis gases in a smoke layer produced by different panel types when exposed to higher temperatures?
The answers to this question will be generated by the results of the mass-loss experiments, combined with Ozone simulation in an Excel workbook. The Excel workbook will generate the total influenced area and the total mass-loss of a building in the pre-flashover phase. Furthermore it will give insight in the amount of pyrolysis gases that are possible present in the smoke layer. A model of the data flow is explained in chapter 4.
1.6 Limitations
During this research the pre-flashover will be addressed, meaning that the results of this research will not contain any data about the behavior of insulation materials above 400 degrees Celsius. And results of this research can not be used to generate predictions for fully developed fires. Since materials will behave different at higher temperatures.
Data has been generated till temperatures up to 400 degrees Celsius due to the limited capacity of the furnace, which is used to test the sandwich panel samples.
The amount of different tested products has been limited to four, respectively: one PUR panel type, one PIR panel type, one stone wool wall panel type, and stone wool roof panel type. A wider variety of tested materials would give a better overview of the current situation.
This research only focusses on the steel sandwich panel market this market is relatively small in the Netherlands, a further research in the insulations materials applied on buildings build according to the steel deck principle will address a lager amount of buildings.
2 Experimental Test Setup
The aim of this experiment is to determine the mass-loss of the sandwich panels, constructed on standard buildings. This means that during this test the sandwich panels will be heated on one side, to simulate the influence of a hot gass-layer during a fire. This test setup has been discussed before in the M3 report.
2.1Test-procedure
The purpose of this test, in which samples of the panels will be tested, is to create an overview of the thermal behavior of the total sandwich construction, by simulating the thermal load of a hot smoke layer.
In this test setup the samples of the panels will be exposed on one side only to approach a situation as realistic as possible. Decreasing the exposed surface area, by exposing only one side of the sample instead of all four sides, can result in a lower mass-loss, in relation to the exposed time.
Samples
The samples which will be tested need to be currently applied in buildings with slightly pitched roofs, and coated as applied in the most common buildings. As an example, a common sandwich panel has a 0.4-0.6 mm steel facing with a foam/stone wool core and a 0.4-0.6 mm steel top facing. The RC-value can vary between 2 and 3.5. The samples will consist out of three types of cores: PIR/PUR and stone wool, with variation in types of facings: as shown below.
- Three samples with joint - Three samples without joint
(regular facings and coating).
PUR
PIR Stone wool
roof
Stone wool wall
Density (Kg/m³) 30 50 100 100
Thermal conductivity λ
(W/(m*K)) 0.0234 0.023 0.041 0.042
Thickness (mm) 100/135 80/115 60/100 100
Rc value (m²K/w) 4.92 4.59
Weight kg/m² 12.9 11.6 17.7 19.12
Fire safety class B‐s2,d0 B‐s2,d0
B‐s1,d0 A1 A1
Figure 2.1: Materials properties, samples
Each sample will be placed in a tray (Figure 2.5), wrapped in ceramic wool on the sides, to minimise the temperature influence from the sides. A secondary effect that the ceramic wool provides is the freedom of expansion of the sample. There will be
two different types of tests. First of all the complete closed (without joints) construction, this will give an overview of the behavior of a sandwich panel under a thermal load. Secondly the joints will be tested as well to determine the influence of this small opening, and the possibility of gas transport trough this joint.
Gases
The ignition of the pyrolysis product/ the out flowing air can indicate the presence of an explosive mixture in the furnace/ hot smoke layer. It will only show the possible flammability of the mixture. Depending on the supplied air, the LEL will or will not be exceeded.
This topic will not further be addressed in this research.
Duration and temperature
During each test the sample will be exposed to a constant temperature varying between 150 and 350 degrees Celsius. The furnace will be pre-heated at the desired temperature before placing the sample in the furnace. The top layer of the furnace will then be quickly ventilated to bring it back to ambient temperature. The exact temperature will be determined on the surface of the sample at the exposed side:
- 10 minutes at 150 degrees.
- 10 minutes at 250 degrees.
- 10 minutes at 350 degrees.
Constant temperatures are used to ensure repeatability and create intervals in the spreadsheet calculations.
It is possible that panels that have been exposed to the weather for a longer time, may have a different properties due to humidity and aging of the core materials. The humidity percentage in the samples might influence the degradation process. This research is based on new/ from the factory delivered sandwich panels, to ensure repeatability and the reliability of the experiment.
Measurement/observations of sample after the test.
After each test the weight loss of each sample will be measured by a scale, after that will be visually determined the degradation of the core.
- Depth of degradation/damage - Delamination of the facing
- Type of degradation, shrinking, blistering, melting, and char forming
- Gas transport.
2.2 Test setup
Introduction
In this paragraph the setup is described in which the samples of both experiments will be tested. Starting with the furnace and heating element, further on the applied monitoring/ measuring instruments.
The furnace
The furnace (Figure 2.2) has an inner space of 30 cm x 30 cm x 60 cm (l*w*h) and is built out of gas concrete blocks. These have been covered with aluminium foil, to decrease the absorption of pyrolised gases by the gas concrete blocks. The heating of the furnace is done by an electric heating element (Figure 2.3) of 2300 Watt. To hold the samples in place, steel L- profiles are mounted to the sides of the furnace (Figure 2.4).
The sample it self will be placed in a steel frame with dimensions 29.8 cm x 29.8 cm made out of L-steel profile 50*50*5mm and a self weight of 4,133 kg as shown in Figure 2.5. This way it can be placed in and out of the furnace in an easy way. The sample will be insulated with ceramic insulation to prevent heat transfer through the sides of the sample. The samples of the LFL experiment will be placed in a bucket and hung onto the scale which is placed on the top of the furnace. The placing of the cover of the furnace is the main difference between the LFL and Mass-loss experiment. To create a realistic situation for the mass-loss experiment the top part of the furnace needs to be open, so the “outside” of the product is exposed to ambient temperatures.
To minimise the risk of a smoke-gas explosion within the furnace, an air supply is integrated in the furnace. In this way the furnace can be ventilated, in a controlled way at the demanded air supply.
The density and pressure of the supplied air will be controlled by the mass flow controller, of which the inlet is positioned approximately 15 cm from the bottom, left hand wall and right hand wall. A bit higher is the fan positioned, this fan is meant to create an homogeneous mixture in the furnace. This fan is powered by a 12 volt power supply. Figure 2.6 shows the “technical” drawing of the furnace
Heat source
The heating of the samples will be done by an electronic heating element, to eliminate the influence of possible unburned fuels when using a gas burner.
The used heating element is made out of massive messing plates. In which three elements of 800 watt are placed. This element is placed in the middle of the base of the furnace 3 cm from the bottom and 0,5 cm from the front and back wall, approximately,6
Figure 2.2: Test setup
Figure 2.3: Heating element
Figure 2.4: Rails to support the sample tray’s
Figure 2.5: Sample placed in the tray, without ceramic wool
- Thermocouple K4: Standard thermocouple type K, positioned the middle of the core material of the tested sample, at 3 cm from the exposed surface.
This thermocouple will monitor rise in temperature of the core material which can be compared with the simulations as shown in paragraph 3.1.4
- Thermocouple K5: Plate-thermocouple type K brazed on a thin metal sheet (3*3 cm), which is positioned at an height of 30 cm, 12 cm off set from the right-hand wall and 1 cm from the front wall.
The distance between the heating element and the thermocouple is 24 cm. This thermocouple is not logged, but used as a thermostat. Due to the fact that using the logger for both situations might disturb the signal. The standard deviation between K3 and K5 is 3 °C, this deviation might have been caused by the settings of the thermostat.
Both plate-thermocouples have not been covered at the non exposed surface, meaning that they look both ways, and therefore follow a more realistic pattern of the actual temperature of the exposed steel facing.
Scale
The scale, that will be used for determining the weight of the sample as well in the mass-loss experiment as in the LFL experiment, is a Mettler Toledo PB3001 produced in 1997. This device has an accuracy of 0.1 gram. Starting at 5 grams and a maximum weight of 3100 grams.
Heat transfer analyzer
The Isomet Heat Transfer Analyzer, will be used to determine the actual thermal conductivity of the insulation materials. Samples exposed to 350 degrees Celsius will also be analyzed on the area that is visibly influenced by the heats source after the test, regarding the 1,5 cm offset from the sides of the sample.
Vernier caliper
This device has been used to measure the intruding depth of the samples. An exact part of 400 cm² has been exposed directly to the heat source (20*20 cm).
The sample needs to be cut in half, the measurement with the vernier caliper will start on the left hands side at the first position that has been exposed. The depth is measured by the stem of the caliper, each 5 cm, till the depth that the core material is deformed and/or discolored.
cm from the right and left had wall. By placing the heating element on this position the sample will be equally exposed to the radiation. The heating element will be connected to a plate-thermocouple that has been placed at the front side of the furnace, just below the surface of the frame in which the sample is placed.
This way the temperature at the height of the sample is determined without blocking a part of the sample.
This method will improve the speed of switching samples, and the temperature will be measured at the exact same spot during all the tests. The heat element will be controlled through the sheet thermocouple which is attached to the self adjusting temperature controller type West 6100. To calibrate this setup a baseline measurement is necessary. This is done by heating the furnace when it is empty, for the LFL test and when a sample of the tested product is placed.
For the Mass-loss test, if a similar RC-value is used to calibrate the furnace. The best result will be achieved.
2.3 Used instruments
In this setup five thermocouples are used of, two different types. Thermocouple K1, K2, and K4 are all standard thermocouples. Thermocouple K3 and K5 are both modified to a plate thermocouple, so they can monitor the surface temperature of the sample, which is expected to diver from the air temperature, due to the chosen heat source. Besides differences in temperature, the plate-thermocouples are expected to have a longer reaction time and they will therefore give a more stable temperature reproduction.
All thermocouples (K1, K2, K3, K4) except for K5 are linked to a data logger (squirrel data logger Grant 2010 series). Thermocouple K5 is linked to the temperature controller (West 6001) together with the dimmer pack dmx152X.
Thermocouples:
- Thermocouple K1: Standard thermocouple type K, positioned at a height of 25 cm, 3.5 cm off set from (right-hand wall) and 10 cm entering the furnace.
- Thermocouple K2: Standard thermocouple type K, positioned at a height of 49 cm, 15 cm off set from the right- had wall and 4 cm entering the furnace.
- Thermocouple K3: Plate-thermocouple type K brazed on a thin metal sheet (3*3 cm). Positioned at the height of 30 cm, 15 cm off set from the right- hand wall and 1 cm from the front wall. The distance between the heating element and the thermocouple is 24 cm.
250
490110600100
35
100 150 150 100 600
100 100
350490110600100
100 285 100
600
100 100
K3 K2
K1
K2
K1
350600100
100 500
600
100 100
250
K3
100 150 150 100
K4
250
K3 K1 K2 300
100 100
260
100 40 100
100 150 150 100
300100100 26510010035100200100100 A
B
C D
Side A
Side C
Section 1-1
1
1
K2
K1
350
600100 250 250
490110
260
100 40 100
100 150 150 100 300
100 100
2 2
Section 2-2
Test sample
Electric heating 2100W
1 2 3 4
Door Steel profile
Parts list
2 2
3
3
4 4
1 1
Figure 2.6: Technical drawing test setup
2.4 Future possibilities test setup
The test setup as explained in this chapter, is suitable for testing building material samples, that need to be exposed to heat, at one single side up to temperatures of 400 degrees Celsius. This testing method gives insight in the behavior of materials exposed to temperatures in the pre-flashover phase. And therefore it might be a welcome addition to the regular fire tests that are applied on building products. Due to the bigger sample size it is possible to test assembled constructions, and composite building products. As mentioned before, EPS and other tested materials are also applied on building build according to the steel-deck principle Figure 2.7. The current test setup is also suitable for testing these kind of building elements.
Figure 2.7 Detail 3.1: Taken from SBR [22]
2.5 Validation test setup
This section will describe the behavior of the thermocouples of the test setup. During the experiments there will be four thermocouples monitoring temperature. Three different kinds of tests have been done to create insight in the behavior of the test setup, regarding the maximum temperature regulation, and self-regulating functions. All test- runs have been done without sample materials.
2.5.1Test-runs
The first test-run was used to determine the maximum temperature of the furnace, with the current heating element which has a maximum power of 2400 watt.
The total running time 2 hours and 23 minutes, with a time interval of 10 sec. during this test the fan has been running a few rotations a minute. The rise of temperature has been examined on the different time intervals while using the full power of the heating element. The time interval from ambient temperature to 150 °C shows an increase of temperature of 3.6
°C each minute. The of 3.1 °C each minute The time interval 250 to 360 has a temperature increase of 1.5
°C. Concluding that the heating process when the heat element is on full power will decrease as the temperature rises.
The heating process as shown belongs to a calibrated temperature controller but in a controlled environment, and a set point of 150°C. The reaction of the plate thermocouples is a bit slower but more stable than the normal thermocouples, as expected.
During test run 8 the heating process from ambient temperature to a stable temperature of 250 °C was logged this stable temperature is reached after 80 minutes. During this test the same response of the thermocouples and the self-controlling thermostat has been noticed.
Using this self-controlling thermostat it is possible to create very stable temperatures on the exact desired temperatures.
Test run 7B starts at a stable temperature of 150 °C during the test run the event of placing a tray with a sample as used in the mass-loss experiment has been simulated. Figure 2.8 shows the placing of a tray with low density mineral wool of 5 cm thickness, which will be exposed for 10 minutes. The top of the furnace is not removed during this simulations.
A quick recovery of temperature is shown, the temperature even exceeds the set point of 150
degrees, and levels out at 158 °C. The air temperature rises more in the beginning and will stabilize at 165
°C. The difference of these two thermocouples can be explained by the difference in height in position.
The rise in temperature in the furnace is caused by the decrease of 50 % in volume of the furnace, when a sample is placed.
70 90 110 130 150 170 190
10 40 70 100 130 160 190 220 250 280 310 340 370 400 430 460 490 520 550 580 610 640 670 700 730
Temperature (oC)
Time (s)
Temperature testrun
K1, air bottom K2, air top K3, plate rails
After finishing the 10 minute test the sample needs to be removed from the furnace, and the temperature in the furnace shall drop by the opening of the door.
The temperature recovery process is shown in Figure 2.9 It shows the time the furnace needs to recover its set point. This event takes approximately 10 minutes, but can differ in other weather conditions.
70 80 90 100 110 120 130 140 150 160 170
670 700 730 760 790 820 850 880 910 940 970 1000 1030 1060 1090 1120 1150 1180 1210 1240 1270 1300
Temperature (oC)
Time (s)
Temperature testrun
K1, air bottom K2, air top K3, plate rails
Figure 2.10 shows the behavior of the temperature while placing the tray with sample after the recovery of the furnace. There is a drop of 20 degrees in the beginning. After three minutes the temperature stabilizes 15 degrees above set point. The quick recovery and stable temperatures are caused by the thermal mass of the furnace, a quick adjustment in temperature is therefore not possible.
Figure 2.8: Placing a test sample 7B1
Figure 2.9: Recovery of the furnace temperature 7B2
0 20 40 60 80 100 120 140 160 180
1270 1300 1330 1360 1390 1420 1450 1480 1510 1540 1570 1600 1630 1660 1690 1720 1750 1780 1810 1840 1870 1900 1930 1960 1990
Temperatuur (oC)
Time (s)
Temperature testrun
K1, air bottom K2, air top K3, plate rails
2.5.2 Sample placing procedure
A step to step explanation of the placing of the sample 1. The furnace is closed, lid is placed on the furnace and the tray filled with low density mineral wool is placed on the sample position.
2. The furnace will be heated respectively 10, 20 and 30 degrees above the desired temperature. (150, 250 or 350 degrees Celsius)
3. The weight of the sample will be determined, secondly the weight included sample will be placed in the tray, surrounded by ceramic wool.
4. When the furnace reaches the temperature the logger will be turned on, the tray with the sample will be placed in the furnace. The top will be removed, and the timer is set on 10 min.
5. After 10 minutes the trays will be switched again, and the lid is replaced on top of the furnace.
6. The process restarts again.
Some remarks, to operate the furnace:
- To establish a representative test, the tray needs to be heated up till furnace temperature before the first test;
- The furnace needs to have a higher temperature as desired before placing the sample;
- Before placing the sample it is useful to raise the set point so the heating element starts heating up. This way the temperature drop will be corrected faster. This element has a lot of mass and reacts slow;
- Temperature drops at 150 degrees Celsius are approximately 10 degrees;
- Temperature drops at 250 degrees Celsius are approximately 25 degrees;
- Temperature drops at 350 degrees Celsius are approximately 30 degrees.
2.5.3. Behaviour in outside conditions
The setup will be used outside, since there is no suitable smoke extracting system available at the test location. Experiments will be carried out in dry weather. Nevertheless weather influences such as temperature, and wind speed might influence the behavior of the thermocouples and the increase of the heat in the furnace. This is important to take into account when generating the results. The same tests at different days may give different results.
The thermocouples exposed to the outside conditions such as K2 and K3, may vary by the influence of wind while opening the door.
The temperature of the furnace is mainly influenced by wind and while the trays with the samples are switched. The switching time has a huge influence at the furnace temperature, especially at higher temperatures.
Figure 2.10: Placing the second test sample 7B3
